]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
28 years agoAdd ATAPI_STATIC so that the ATAPI cdroms work correctly again
Jordan K. Hubbard [Mon, 29 Apr 1996 20:03:41 +0000 (20:03 +0000)]
Add ATAPI_STATIC so that the ATAPI cdroms work correctly again
under -current.
Submitted-By: Serge Vakulenko <vak@cronyx.ru>
28 years agoAdd an extra line to the floppies target which will make it refresh
Jordan K. Hubbard [Mon, 29 Apr 1996 19:38:36 +0000 (19:38 +0000)]
Add an extra line to the floppies target which will make it refresh
the distribution area's floppy images.

28 years agoReshuffle the installation routines a little for greater clarify.
Jordan K. Hubbard [Mon, 29 Apr 1996 19:34:25 +0000 (19:34 +0000)]
Reshuffle the installation routines a little for greater clarify.

28 years agoYikes - bogon crept into the NFS setup menu.
Jordan K. Hubbard [Mon, 29 Apr 1996 18:06:08 +0000 (18:06 +0000)]
Yikes - bogon crept into the NFS setup menu.

28 years agoFix a small display bogon with the progress indicators.
Jordan K. Hubbard [Mon, 29 Apr 1996 17:59:07 +0000 (17:59 +0000)]
Fix a small display bogon with the progress indicators.

28 years agoFixed two typos in the comment.
Sujal Patel [Mon, 29 Apr 1996 15:07:59 +0000 (15:07 +0000)]
Fixed two typos in the comment.

Pointed out by: davidg

28 years agoVery minor tweak:
Bill Paul [Mon, 29 Apr 1996 14:48:47 +0000 (14:48 +0000)]
Very minor tweak:

In __initdb(), a failure to open the local password database is supposed
to result in a warning message being syslog()ed. This warning is only
supposed to be generated as long as the 'warned' flag hasn't been yet;
once the warning is generated, the flag should be set so that the message
is only syslog()ed once. However, while the state of the flag is checked
properly, the flag's state is never changed, so you always get multiple
warnings instead of just one.

Pointed out by: Peter Wemm

28 years agoAdd some more debugging statements.
Jordan K. Hubbard [Mon, 29 Apr 1996 06:47:10 +0000 (06:47 +0000)]
Add some more debugging statements.

28 years agoFix mis-declared static arrays that made sysinstall SEGV in
Jordan K. Hubbard [Mon, 29 Apr 1996 06:45:33 +0000 (06:45 +0000)]
Fix mis-declared static arrays that made sysinstall SEGV in
Set_Boot_Blocks().  Boy, this one had me tearing my hair out!
I hate how the loader distinguishes between `extern char *foo' and
`extern char foo[]' sometimes! :-)

28 years agoFix embarassing mind-o: transient RPC program numbers run from
Bill Paul [Mon, 29 Apr 1996 05:24:26 +0000 (05:24 +0000)]
Fix embarassing mind-o: transient RPC program numbers run from
0x40000000 to 0x5FFFFFFF, not 0x4000000 to 0x5FFFFFF. *sigh*

28 years agoFix some display bogons in the label editor.
Jordan K. Hubbard [Mon, 29 Apr 1996 05:09:23 +0000 (05:09 +0000)]
Fix some display bogons in the label editor.

28 years agoMove some warn()'s into DEBUG space since I don't need them coming
Jordan K. Hubbard [Mon, 29 Apr 1996 05:03:02 +0000 (05:03 +0000)]
Move some warn()'s into DEBUG space since I don't need them coming
out in my curses interfaces and spamming my screen.

28 years agoDon't stop /etc/sysconfig if we're `faking'.
Jordan K. Hubbard [Sun, 28 Apr 1996 22:54:21 +0000 (22:54 +0000)]
Don't stop /etc/sysconfig if we're `faking'.
More asthetic adjustments to menus.
Improve both the display efficiency and usability of the label
screen - it's now possible to label quite a few disks in the label editor.

28 years agoAdd a new -fake argument to make sysinstall not actually change things while
Jordan K. Hubbard [Sun, 28 Apr 1996 20:54:11 +0000 (20:54 +0000)]
Add a new -fake argument to make sysinstall not actually change things while
I'm testing it.

28 years agoBring back the AHC_TAGENABLE compile option and document it. You must
Justin T. Gibbs [Sun, 28 Apr 1996 19:43:10 +0000 (19:43 +0000)]
Bring back the AHC_TAGENABLE compile option and document it.  You must
use this option to get tagged queuing.  Just like the old days.

28 years agoFix the abort code in the BUS DEVICE RESET case. Needed to set cmdlen to 0.
Justin T. Gibbs [Sun, 28 Apr 1996 19:21:20 +0000 (19:21 +0000)]
Fix the abort code in the BUS DEVICE RESET case.  Needed to set cmdlen to 0.

Ensure that queued commands are not touched by the abort code by setting
the SCB status to indicate what queue it is in.

Fix deadlocks when using SCB paging by using SCBs from the assigned_scbs
queue or an SCB that completed during the same interrupt if needed.

Don't ever use insl to pull SCBs from any of the controllers.  You can
only do 8bit PIO reads.  This only affected SCB paging.

With this checkin, SCB paging works quite a bit better, but I still have
some problems with it that may be caused by a firmware problem in my
PD1800s.  It seems that using a tag number higher than the maximum number
of tags allowed by the device, confuses it.  For example, if I queue
two commands, tagged 3 and 36, it never reconnects for tag 36.

28 years agoAdd an entry for xterm with colors. The standard xterm now handles
Joerg Wunsch [Sun, 28 Apr 1996 18:21:01 +0000 (18:21 +0000)]
Add an entry for xterm with colors.  The standard xterm now handles
ANSI colors.

28 years agoFix a bogon that prevented easy-import from inserting the new module
Joerg Wunsch [Sun, 28 Apr 1996 12:32:43 +0000 (12:32 +0000)]
Fix a bogon that prevented easy-import from inserting the new module
name in alphabetical order (broken by my previous commit).

Pointed out by: asami

28 years agoFix some bugs I introduced and some old ones as well.
Poul-Henning Kamp [Sun, 28 Apr 1996 07:14:05 +0000 (07:14 +0000)]
Fix some bugs I introduced and some old ones as well.
Add BDE_DEBUGGER back.
Improve quality of comments.
Thanks Bruce!

Reviewed by: phk
Submitted by: bde

28 years agoUpdate to new SNAP revision information.
Jordan K. Hubbard [Sun, 28 Apr 1996 06:08:41 +0000 (06:08 +0000)]
Update to new SNAP revision information.

28 years agoPerformance enhancements (I hope) and new stuff:
Bill Paul [Sun, 28 Apr 1996 04:38:52 +0000 (04:38 +0000)]
Performance enhancements (I hope) and new stuff:

yp_dblookup.c:

- Implement database handle caching. What this means is that instead
  of opening and closing map databases for each request, we open a
  database and save the handle (and, if requested, the key index)
  in an array. This saves a bit of overhead on things like repeated
  YPPROC_NEXT calls, such as you'd get from getpwent(). Normally,
  each YPPROC_NEXT would require open()ing the database, seeking
  to the location supplied by the caller (which is time consuming with
  hash databases as the R_CURSOR flag doesn't work), reading the
  data, close()ing the database and then shipping the data off to
  the caller. The system call overhead is prohibitive, especially
  with very large maps. By caching the handle to an open database,
  we elimitate at least the open()/close() system calls, as well
  as the associated DB setup and tear-down operations, for a large
  percentage of the time. This improves performance substantially at
  the cost of consuming a little more memory than before.

  Note that all the caching support is surrounded by #ifdef DB_CACHE
  so that this same source module can still be used by other programs
  that don't need it.

- Make yp_open_db() call yp_validdomain(). Doing it here saves cycles
  when caching is enabled since a hit on the map cache list by
  definition means that the domain being referenced is valid.

- Also make yp_open_db() check for exhaustion of file descriptors,
  just in case.

yp_server.c:

- Reorganize things a little to take advantage of the database
  handle caching. Add a call to yp_flush_all() in ypproc_clear_2_svc().

- Remove calls to yp_validdomain() from some of the service procedures.
  yp_validdomain() is called inside yp_open_db() now, so procedures that
  call into the database package don't need to use yp_validdomain()
  themselves.

- Fix a bogosity in ypproc_maplist_2_svc(): don't summarily initiallize
  the result.maps pointer to NULL. This causes yp_maplist_free()
  to fail and leaks memory.

- Make ypproc_master_2_svc() copy the string it gets from the database
  package into a private static buffer before trying to NUL terminate it.
  This is necessary with the DB handle caching: stuffing a NUL into the
  data returned by DB package will goof it up internally.

yp_main.c:

- Stuff for DB handle caching: call yp_init_dbs() to clear the
  handle array and add call to yp_flush_all() to the SIGHUP
  signal handler.

Makefile.yp:

- Reorganize to deal with database caching. yp_mkdb(8) can now be used
  to send a YPPROC_CLEAR signal to ypserv(8). Call it after each map
  is created to refresh ypserv's cache.

- Add support for mail.alias map.
  Contributed by Mike Murphy (mrm@sceard.com).

- Make default location for the netgroups source file be /var/yp/netgroup
  instead of /etc/netgroup.

mkaliases:

- New file: script to generate mail.alias map.
  Contributed by Mike Murphy (mrm@sceard.com).

Makefile:

- Install Makefile.yp as /var/yp/Makefile.dist and link it to
  /var/yp/Makefile only if /var/yp/Makefile doesn't already exist.
  Suggested by Peter Wemm.

- Install new mkaliases script in /usr/libexec along with mknetid.

- Use somewhat saner approach to generating rpcgen-dependent files
  as suggested by Garrett Wollman.

28 years agoImport the new yp_mkdb. This one does what the old one does, but is
Bill Paul [Sun, 28 Apr 1996 04:16:06 +0000 (04:16 +0000)]
Import the new yp_mkdb. This one does what the old one does, but is
hopefully less grotty. Differences include:

- An extra flag, -c, for sending a YPPROC_CLEAR to ypserv(8) (which
  will become useful once I merge in the DB handle caching mods I've
  been working on).
- Checks for and signals duplicate keys.
- Handles line continuations.

As of now, there is no more GNU YP code in the tree.

28 years agoToss the GNU yp_mkdb into the attic.
Bill Paul [Sun, 28 Apr 1996 04:08:22 +0000 (04:08 +0000)]
Toss the GNU yp_mkdb into the attic.

28 years agoRemove yp_mkdb; it's going away.
Bill Paul [Sun, 28 Apr 1996 04:06:06 +0000 (04:06 +0000)]
Remove yp_mkdb; it's going away.

Uhm, guys? This Makefile doesn't build anything anymore. The isdn
stuff under here isn't compiled. Is this normal?

28 years agoSmall tweak to yp_put_record(): call the DB put routine with the
Bill Paul [Sun, 28 Apr 1996 03:59:56 +0000 (03:59 +0000)]
Small tweak to yp_put_record(): call the DB put routine with the
R_NOOVERWRITE flag and process return codes so that we can tell the
difference between a failure due to a duplicate database entry and
failure due to some other error.

28 years agoFix an ancient index list bug.
Jordan K. Hubbard [Sun, 28 Apr 1996 03:27:26 +0000 (03:27 +0000)]
Fix an ancient index list bug.
Make FTP re-initializion work.
Fix fix fix.

28 years agoFix the OK/Cancel accellerator bogon I introduced a few days back.
Jordan K. Hubbard [Sun, 28 Apr 1996 02:13:07 +0000 (02:13 +0000)]
Fix the OK/Cancel accellerator bogon I introduced a few days back.

28 years agoMove some of functions around in order to make this stuff easier
Jordan K. Hubbard [Sun, 28 Apr 1996 01:07:27 +0000 (01:07 +0000)]
Move some of functions around in order to make this stuff easier
to work on.

28 years agoBump snapshot release date - looks like I was off by a couple
Jordan K. Hubbard [Sun, 28 Apr 1996 00:37:38 +0000 (00:37 +0000)]
Bump snapshot release date - looks like I was off by a couple
of days!

More display fixes to sysinstall and assorted tweaking.

28 years agoFix a minor grammatical error (reselled -> resold) I just noticed.
Jordan K. Hubbard [Sat, 27 Apr 1996 18:36:02 +0000 (18:36 +0000)]
Fix a minor grammatical error (reselled -> resold) I just noticed.

28 years agoinet_ntoa buffer was evaluated twice in log_in_vain, fix it.
Andrey A. Chernov [Sat, 27 Apr 1996 18:19:12 +0000 (18:19 +0000)]
inet_ntoa buffer was evaluated twice in log_in_vain, fix it.
Thanx to: jdp

28 years agoMake `install' depend on `kernel'. Some idiot tried to install a
Joerg Wunsch [Sat, 27 Apr 1996 07:30:49 +0000 (07:30 +0000)]
Make `install' depend on `kernel'.  Some idiot tried to install a
non-built kernel and complained in Usenet... :-)

28 years agoAdd a missing helpfile for fixit.
Jordan K. Hubbard [Sat, 27 Apr 1996 07:18:59 +0000 (07:18 +0000)]
Add a missing helpfile for fixit.

28 years agoFix what I just broke with DITEM_LEAVE_MENU
Jordan K. Hubbard [Sat, 27 Apr 1996 07:09:41 +0000 (07:09 +0000)]
Fix what I just broke with DITEM_LEAVE_MENU

28 years agoMore misc menu hackery.
Jordan K. Hubbard [Sat, 27 Apr 1996 07:04:12 +0000 (07:04 +0000)]
More misc menu hackery.

28 years agoFix a minor typo.
Jordan K. Hubbard [Sat, 27 Apr 1996 04:57:43 +0000 (04:57 +0000)]
Fix a minor typo.

28 years agoFix a bogon with ok/cancel callback usage.
Jordan K. Hubbard [Sat, 27 Apr 1996 01:52:15 +0000 (01:52 +0000)]
Fix a bogon with ok/cancel callback usage.

28 years agoUpdate more docs to match -current reality.
Jordan K. Hubbard [Sat, 27 Apr 1996 01:34:47 +0000 (01:34 +0000)]
Update more docs to match -current reality.

28 years agoMake default /etc/motd (which I will include in the installation) a little
Jordan K. Hubbard [Sat, 27 Apr 1996 00:44:22 +0000 (00:44 +0000)]
Make default /etc/motd (which I will include in the installation) a little
more helpful and explanatory to new users.  Experienced users always blow
away their default motd's on the first day anyway. :-)

28 years agoFixed a bug introoduced in the previous change. ISA device memory was
Bruce Evans [Fri, 26 Apr 1996 22:26:45 +0000 (22:26 +0000)]
Fixed a bug introoduced in the previous change.  ISA device memory was
mapped to semi-random place(s) depending on the content(s) of physical
address 0xA0000.  This was fatal at least on my system with a some
memory-mapped devices.  Console syscons somehow wasn't affected.  It
bogusly hardcodes the address.  Sigh.

28 years agoImplement keyinfo(1) as a setuid Perl script, so /etc/skeykeys can
Joerg Wunsch [Fri, 26 Apr 1996 21:35:11 +0000 (21:35 +0000)]
Implement keyinfo(1) as a setuid Perl script, so /etc/skeykeys can
be made mode 0600.

Pointed out by: kieber@sax.de (Ulf Kieber)
Reviewed by: pst

28 years ago/etc/skeykeys was basically suffering from the same vulnerability
Joerg Wunsch [Fri, 26 Apr 1996 21:33:18 +0000 (21:33 +0000)]
/etc/skeykeys was basically suffering from the same vulnerability
as any non-shadowed /etc/passwd.  Ironically, all programs using S/Key
have already been setuid root except keyinfo(1).

This modification creates /etc/skeykeys with mode 0600 to prevent it
from being examined by ordinary users.

28 years agoRemove the comments that suggest people not turn
Mike Pritchard [Fri, 26 Apr 1996 20:18:58 +0000 (20:18 +0000)]
Remove the comments that suggest people not turn
on accounting and quotas because they don't work.
They seem to work just fine, and the comments generate
too many questions.

28 years agoRemove libdisk from the old location.
Jordan K. Hubbard [Fri, 26 Apr 1996 19:27:47 +0000 (19:27 +0000)]
Remove libdisk from the old location.

28 years agochanged a strchr to a strrchr so that modules with a `.' in the path
Eric L. Hernes [Fri, 26 Apr 1996 18:39:48 +0000 (18:39 +0000)]
changed a strchr to a strrchr so that modules with a `.' in the path
can be loaded (e.g. ./mymod.o)

28 years agoDelete #ifdef notdef blocks containing old method of srtt calculation.
Garrett Wollman [Fri, 26 Apr 1996 18:32:58 +0000 (18:32 +0000)]
Delete #ifdef notdef blocks containing old method of srtt calculation.

Requested by: davidg

28 years agoDelete #if 0 block containing remnants of pre-MTU discovery rmx_mtu
Garrett Wollman [Fri, 26 Apr 1996 18:31:41 +0000 (18:31 +0000)]
Delete #if 0 block containing remnants of pre-MTU discovery rmx_mtu
initialization.

28 years agoDelete #if 0 block containing unused definitions for ARPANET/DDN IMP
Garrett Wollman [Fri, 26 Apr 1996 18:30:52 +0000 (18:30 +0000)]
Delete #if 0 block containing unused definitions for ARPANET/DDN IMP
and HYPERchannel link layers.

28 years agoFix display refresh bug in displayHelp()
Jordan K. Hubbard [Fri, 26 Apr 1996 18:19:38 +0000 (18:19 +0000)]
Fix display refresh bug in displayHelp()
Go back to Normal menus for Media and FTP menus rather than
radio menus - the difference in behavior is confusing and sort
of detracts from the added (small) advantage of seeing what you last
picked.

28 years agoA significant debogofication of locore.s. I havn't found any actualy
Poul-Henning Kamp [Fri, 26 Apr 1996 13:47:39 +0000 (13:47 +0000)]
A significant debogofication of locore.s.  I havn't found any actualy
bugs, but it is a lot easier to navigate this twisted code now.

28 years agoAdd CFLAGS=${CFLAGS} to MAKE_ENV, otherwise gmake won't pick it up.
Satoshi Asami [Fri, 26 Apr 1996 10:09:06 +0000 (10:09 +0000)]
Add CFLAGS=${CFLAGS} to MAKE_ENV, otherwise gmake won't pick it up.

28 years agoRemoved options MAXCONS & HARDFONT, they are no longer in use
Søren Schmidt [Fri, 26 Apr 1996 06:45:41 +0000 (06:45 +0000)]
Removed options MAXCONS & HARDFONT, they are no longer in use
(and havn't been for long, sigh)

28 years agoAll hail Mike Pritchard - the new Docmaster!
Jordan K. Hubbard [Thu, 25 Apr 1996 20:07:28 +0000 (20:07 +0000)]
All hail Mike Pritchard - the new Docmaster!
Nominated by: jfieber
Dragooned by: jkh
Foolishly Accepted by: mpp

28 years agoMight as well update the release notes while I'm here!
Jordan K. Hubbard [Thu, 25 Apr 1996 19:04:45 +0000 (19:04 +0000)]
Might as well update the release notes while I'm here!

28 years agoThese docs were in fair need of updating.
Jordan K. Hubbard [Thu, 25 Apr 1996 18:40:02 +0000 (18:40 +0000)]
These docs were in fair need of updating.

28 years agoOops, restore missing piece of previous commit.
Andrey A. Chernov [Thu, 25 Apr 1996 18:14:11 +0000 (18:14 +0000)]
Oops, restore missing piece of previous commit.

28 years agoAdd a place-holder message for the CDROM fixit feature (which I'm not going
Jordan K. Hubbard [Thu, 25 Apr 1996 18:00:28 +0000 (18:00 +0000)]
Add a place-holder message for the CDROM fixit feature (which I'm not going
to add in time for the next SNAP).

28 years agoQuite a few changes:
Jordan K. Hubbard [Thu, 25 Apr 1996 17:31:27 +0000 (17:31 +0000)]
Quite a few changes:

1. Fix the last display bugs (I hope) by use of dialog rebuilds at stategic
   points.
2. Clean up the distributions menus so that everybody (that's reasonable)
   has All and Clear options for setting/clearing things en-masse.
3. Various attempts at display optimization.
4. Change the wording of the `Don't use Write!' dialogs to make them more
   explicitly define when and when not to use the option.

28 years agoMore custom hackery to deal with issues discovered in sysinstall.
Jordan K. Hubbard [Thu, 25 Apr 1996 17:27:18 +0000 (17:27 +0000)]
More custom hackery to deal with issues discovered in sysinstall.

28 years agoTreat ^end<EOF> as valid terminator (like ^end\n), some encoders (f.e. Eudora)
Andrey A. Chernov [Thu, 25 Apr 1996 14:49:40 +0000 (14:49 +0000)]
Treat ^end<EOF> as valid terminator (like ^end\n), some encoders (f.e. Eudora)
produce such things.

28 years agoAllow some post-fire actions to be combined in ways that weren't
Jordan K. Hubbard [Thu, 25 Apr 1996 13:34:37 +0000 (13:34 +0000)]
Allow some post-fire actions to be combined in ways that weren't
possible before.

28 years agoFix cpu_fork for real.
Poul-Henning Kamp [Thu, 25 Apr 1996 06:20:19 +0000 (06:20 +0000)]
Fix cpu_fork for real.

Suggested by:  bde

28 years agoFix logic bug in pci bridge code. For a PCI-PCI bridge, secondary
Satoshi Asami [Thu, 25 Apr 1996 06:04:27 +0000 (06:04 +0000)]
Fix logic bug in pci bridge code.  For a PCI-PCI bridge, secondary
should be <= than subordinate, not the other way around.

They are both true if the bridge is not cascaded (i.e., twin-channel
scsi/e-net adapters won't be affected by this bug), which is probably why
it was unnoticed until today.

28 years agoRegardless of whether or not the check for IPv4 is useful, we certainly
David Greenman [Thu, 25 Apr 1996 02:34:37 +0000 (02:34 +0000)]
Regardless of whether or not the check for IPv4 is useful, we certainly
don't need to assign the "ip" pointer twice.

28 years agoalloca() was being called with highly bogus arguments due to brain
Jordan K. Hubbard [Thu, 25 Apr 1996 01:50:18 +0000 (01:50 +0000)]
alloca() was being called with highly bogus arguments due to brain
fade.  Fixed.

28 years agoThe traceon and traceoff directives aren't in this version ncurses.
Jordan K. Hubbard [Thu, 25 Apr 1996 01:18:30 +0000 (01:18 +0000)]
The traceon and traceoff directives aren't in this version ncurses.
Guess nobody's built these tests for quite awhile!

28 years agoWhen a local symbol that would normally be eliminated by "-x" or
John Polstra [Wed, 24 Apr 1996 23:31:08 +0000 (23:31 +0000)]
When a local symbol that would normally be eliminated by "-x" or
"-X" must survive, because of references from relocations, don't
qualify the symbol name with the name of the input file.  This
saves some string space.  It makes libc_pic.a about 2.4% smaller.

Adapted from a suggestion by Bruce Evans.

28 years agoChange "Found end of tape. Load next tape ..." messages to say
Jordan K. Hubbard [Wed, 24 Apr 1996 21:20:17 +0000 (21:20 +0000)]
Change "Found end of tape.  Load next tape ..." messages to say
"volume" instead of tape.  Running cpio off of /dev/fd0 and having
it say "give me the next tape" is kind of ludicrous.. :-)

28 years agoClose a security hole in sliplogin.
Paul Traina [Wed, 24 Apr 1996 20:18:25 +0000 (20:18 +0000)]
Close a security hole in sliplogin.
If you use sliplogin as a user shell (in /etc/passwd) upgrade to this version.
Reviewed by: bde, peter
Submitted by: AUS CERT
Obtained from: Linux sliplogin-2.02

28 years agoReject all IP versions but 4.
Poul-Henning Kamp [Wed, 24 Apr 1996 15:44:34 +0000 (15:44 +0000)]
Reject all IP versions but 4.

28 years agoAdd missing "int" to static var.
Satoshi Asami [Wed, 24 Apr 1996 09:42:22 +0000 (09:42 +0000)]
Add missing "int" to static var.

28 years agoThis fixes kmem_malloc/kmem_free (and malloc/free of objects of > 8K).
John Dyson [Wed, 24 Apr 1996 04:16:45 +0000 (04:16 +0000)]
This fixes kmem_malloc/kmem_free (and malloc/free of objects of > 8K).
A page index was calculated incorrectly in vm_kern, and vm_object_page_remove
removed pages that should not have been.

28 years agoFix reverse logic bug which caused the ethernet configuration to never work.
Nate Williams [Wed, 24 Apr 1996 00:04:10 +0000 (00:04 +0000)]
Fix reverse logic bug which caused the ethernet configuration to never work.

28 years ago- This code adds some backwards compatability hacks for buggy APM BIOS
Nate Williams [Tue, 23 Apr 1996 19:59:14 +0000 (19:59 +0000)]
- This code adds some backwards compatability hacks for buggy APM BIOS
  implementations, and synchronizes us with the Nomad's latest code.

This code is based on the Nomad code, but heavily hacked by me.

Reviewed by: phk
Submitted by: the 'Nomads'

28 years agoFormat the already applied PCCARD additions to use KNF form. No
Nate Williams [Tue, 23 Apr 1996 18:36:56 +0000 (18:36 +0000)]
Format the already applied PCCARD additions to use KNF form.  No
functional changes.

28 years ago- Make pccardd release it's allocated io space and sets the mem.cardaddr
Nate Williams [Tue, 23 Apr 1996 16:46:48 +0000 (16:46 +0000)]
- Make pccardd release it's allocated io space and sets the mem.cardaddr
  even if DEBUG is not set.
- #ifdef protect DEBUG printf's

Submitted by: sos

28 years agoAdds the APM hooks into the generic pccard kernel files. With this code
Nate Williams [Tue, 23 Apr 1996 16:03:08 +0000 (16:03 +0000)]
Adds the APM hooks into the generic pccard kernel files.  With this code
in place device drivers can now register power-down/power-up routines so
that we can use common routines to power-up/power-down cards for
insert/removals, suspend/resume, etc..

Reviewed by: phk
Submitted by: the 'Nomads'

28 years agoAdd support for SOCKS if enabled.
Paul Traina [Tue, 23 Apr 1996 06:01:10 +0000 (06:01 +0000)]
Add support for SOCKS if enabled.
Remove a useless bind() call (why did that ever get in there?)

28 years agoAdd support for socks
Paul Traina [Tue, 23 Apr 1996 05:18:43 +0000 (05:18 +0000)]
Add support for socks

28 years agoFix second instance of the inb type when getting the current sequencer address.
Justin T. Gibbs [Tue, 23 Apr 1996 04:47:02 +0000 (04:47 +0000)]
Fix second instance of the inb type when getting the current sequencer address.

Submitted by: Nate Williams <nate@sri.MT.net>

28 years agoFix a typo in the RESTART_SEQUENCER macro that could have caused
Justin T. Gibbs [Tue, 23 Apr 1996 04:22:41 +0000 (04:22 +0000)]
Fix a typo in the RESTART_SEQUENCER macro that could have caused
strange results.  This  bug has been in here for a loooong time.
Many thanks to Pedro Salenbauch for finding this.

Submitted by: pedrosal@nce.ufrj.br (Pedro Salenbauch)

28 years agoBring this into sync.
Jordan K. Hubbard [Tue, 23 Apr 1996 01:29:35 +0000 (01:29 +0000)]
Bring this into sync.

I still have a _very very annoying_ display bug which occurs when a menu
item causes a submenu to be displayed - the screen repaints for the original
menu (which is restored upon return from the submenu) are off by about 4
characters.  I've tried restoring the screen, the cursor position, you name
it - same deal.  Grrrr!  This commit is my first step in trying to get someone
else to help me look into this one since I'm just tearing my hair out at this
point!

28 years agoCosmetic tweaks.
Jordan K. Hubbard [Tue, 23 Apr 1996 01:25:04 +0000 (01:25 +0000)]
Cosmetic tweaks.

28 years agoNew variable NOOBJLINK
Wolfram Schneider [Mon, 22 Apr 1996 23:31:39 +0000 (23:31 +0000)]
New variable NOOBJLINK
Create 'obj' directory in current directory instead
a symbolic link to the 'obj' tree if defined. [not set]

Print a warning if 'obj' tree (/usr/obj) does not exist.

Change default 'obj' directory from ``obj.${MACHINE}'' back to
``obj'', unfortunately many Makefiles are wired with the name ``obj''.

Add some comments for variables and targets.

28 years agoUpdated the laptop 'kludge' file. Hopefully this can go away soon.
Nate Williams [Mon, 22 Apr 1996 22:29:40 +0000 (22:29 +0000)]
Updated the laptop 'kludge' file.  Hopefully this can go away soon.

Submitted by: The 'Nomads'

28 years agotypo shmrsh -> smrsh
Paul Traina [Mon, 22 Apr 1996 20:30:09 +0000 (20:30 +0000)]
typo shmrsh -> smrsh

28 years agoAdd a couple of needed X11 libraries. Discovered when trying to build a
John Polstra [Mon, 22 Apr 1996 20:24:56 +0000 (20:24 +0000)]
Add a couple of needed X11 libraries.  Discovered when trying to build a
static version.

28 years ago- add apm to the GENERIC kernel (disabled by default), and add some comments
Nate Williams [Mon, 22 Apr 1996 19:40:28 +0000 (19:40 +0000)]
- add apm to the GENERIC kernel (disabled by default), and add some comments
  regarding apm to LINT
- Disabled the statistics clock on machines which have an APM BIOS and
  have the options "APM_BROKEN_STATCLOCK" enabled (which is default
  in GENERIC now)
- move around some of the code in clock.c dealing with the rtc to make
  it more obvios the effects of disabling the statistics clock

Reviewed by: bde

28 years agoFix a segfault which occurs when it gets an empty input line.
Sujal Patel [Mon, 22 Apr 1996 18:37:22 +0000 (18:37 +0000)]
Fix a segfault which occurs when it gets an empty input line.
Closes PR bin/1153

Reviewed by: smpatel
Submitted by: Matthew C. Mead (mmead@Glock.COM)

28 years agoThe default number of tags in the SCB paging case was supposed to be
Justin T. Gibbs [Mon, 22 Apr 1996 13:21:40 +0000 (13:21 +0000)]
The default number of tags in the SCB paging case was supposed to be
8 not the 18 I was using during some of my own testing.  Ooops.

For those that want to change the number for experimentation, you can
set the value on line 1553 of this file.

28 years agoThis fixes a troubling oversight in some of the pmap code enhancements.
John Dyson [Mon, 22 Apr 1996 05:23:08 +0000 (05:23 +0000)]
This fixes a troubling oversight in some of the pmap code enhancements.
One of the manifiestations of the problem includes the -4 RSS problem
in ps.

Reviewed by: dyson
Submitted by: Stephen McKay <syssgm@devetir.qld.gov.au>

28 years agoFind official home of smrsh
Paul Traina [Sun, 21 Apr 1996 19:50:46 +0000 (19:50 +0000)]
Find official home of smrsh

28 years agoEnable proper installation of sendmail restricted shell smrsh(8).
Paul Traina [Sun, 21 Apr 1996 19:48:07 +0000 (19:48 +0000)]
Enable proper installation of sendmail restricted shell smrsh(8).

This program is a wrapper for the prog mailer in sendmail.  It does shell
meta character masking and restricts the list of executables to those found
in /usr/libexec/sm.bin.

The default sendmail.cf file does not use this tool, however you can enable
it by either changing /bin/sh to /usr/libexec/smrsh or adding the line
FEATURE(smrsh) into your sendmail .mc file and rebuilding your .cf file.

For more info, RTFMP.

28 years agoAdd /usr/libexec/sm.bin for smrsh(8)
Paul Traina [Sun, 21 Apr 1996 19:40:20 +0000 (19:40 +0000)]
Add /usr/libexec/sm.bin for smrsh(8)

28 years agoAdd "-x" to the linker command for building shared libraries. This
John Polstra [Sun, 21 Apr 1996 17:45:38 +0000 (17:45 +0000)]
Add "-x" to the linker command for building shared libraries.  This
eliminates many local symbols that could not be removed by the "ld -r -x"
steps on the individual object files.  It makes shared libraries
substantially smaller -- almost 11%, in the case of libc.so.3.0.

28 years agoFixed in-line IP header checksumming. It was performed on the wrong header
Bruce Evans [Sun, 21 Apr 1996 13:47:43 +0000 (13:47 +0000)]
Fixed in-line IP header checksumming.  It was performed on the wrong header
in one case.

28 years agoTake out quotes around MOTIFLIB. Clarify comments a bit, most users
Satoshi Asami [Sun, 21 Apr 1996 08:56:32 +0000 (08:56 +0000)]
Take out quotes around MOTIFLIB.  Clarify comments a bit, most users
don't have to touch neither X11BASE nor MOTIFLIB.

Reported by: kmitch@phantasma.bevc.blacksburg.va.us

28 years agoUpdate to face reality. We now support the aic7850, aic7860 and aic7880.
Justin T. Gibbs [Sat, 20 Apr 1996 21:37:49 +0000 (21:37 +0000)]
Update to face reality.  We now support the aic7850, aic7860 and aic7880.

Mention Ultra support and SCB paging.

Remove references to AHC_TAGENABLE.

28 years agoAdd PCI IDs for the aic7860 (basically an aic7850 Ultra) and the aic7855
Justin T. Gibbs [Sat, 20 Apr 1996 21:31:27 +0000 (21:31 +0000)]
Add PCI IDs for the aic7860 (basically an aic7850 Ultra) and the aic7855
(HP motherboard aic7850 controllers).

Properly handle 4 bit controller SCSI IDs.

Update my copyright.

28 years agoImplement SCB paging which allows up to 255 active commands on aic7770
Justin T. Gibbs [Sat, 20 Apr 1996 21:29:27 +0000 (21:29 +0000)]
Implement SCB paging which allows up to 255 active commands on aic7770
(Rev E or greater), aic7850, aic7860, aic7870, and aic7880 controllers.
SCB paging is enabled with the option "AHC_SCBPAGING_ENABLE".  Full
comments on the algorithm are at the top of i386/scsi/aic7xxx.c.

options "AHC_TAGENABLE" and "AHC_QUEUE_FULL" have been removed.  The
default is 4 tags without SCB paging, 8 with.

Clear the SCSIRSTI bit after throwing a bus reset.  Some cards seem to get
confused otherwise.

Handle SCSIRSTI interrupts before checking to see if there is a valid
SCB in use since this can happen. (Clears PR# i386/1123)

Clean up the way we determine the number of SCBs on the card
(courtesy of Dan Eischen).

Guard against attempts to negotiate wide to a narrow controller.

Fix some comments.

Update my copyrights.